73-6
Final Report:
Exhaust Emissions
from a Mercedes-Benz
Diesel Sedan
July 1972
H. Anthony Ashby
Test and Evaluation Branch
Division of Emission Control Technology
Environmental Protection Agency
-------
Background
A Mercedes Diesel sedan, obtained through the courtesy
of Mercedes-Benz of North America, was tested at the Ann Arbor .
Motor Vehicle Emissions Laboratory during the period of
December 1971 through May 1972.
The original objectives of these tests were to develop
a procedure to accurately measure mass emissions from a
Diesel-powered candidate vehicle in the Federal Clean Car "
Incentive Program, and to verify earlier published mass
emissions data on another Mercedes 220 Diesel. Later these
tests became important in the development of an official
Federal Test Procedure for Diesel-powered Light Duty Vehicles,
and the data became of interest in technology assessment efforts
underway in the Division of Emission Control Technology.
Vehicle Description
The car was a Mercedes-Benz 220 Diesel four-door sedan,
with a 134 CID four-cylinder Diesel engine developing 65 SAE
horsepower. Power is transmitted to the rear wheels through
an automatic transmission. The car was tested at a simulated
inertia weight of 3,500 pounds. The first test was at an
odometer reading of 2060 miles and the last at 4145 miles.
Test Program
1. Gaseous Emissions.
A total of 18 cold start 1975 FTP tests (3-bag) will be
reported herein. In addition to these, several tests were
run with the car running at a steady state to investigate the
effects of changing sampling locations and Constant Volume
Sampler (CVS) flow rates.
a. Apparatus
Instrumentation was as specified in the Federal
Register, with exceptions as required by the characteristics
of the Diesel car. Normal CVS bag samples were analyzed
for unburned hydrocarbons (HC) with a Flame lonization
Detector (FID) , for CO and CC>2 with Non-dispersive
Infrared (NDIR) instruments, and for nitrogen oxides (NOx)
with a Chemiluminescence (CL) analyzer. Three different
CO analyzers were used, though not all on the same tests: a
standard cell-length instrument with a range of about 0-2500
ppm, a similar instrument with EPA modifications having a
range of 0-250 ppm, and a long path instrument with a range
of 0-200 ppm.
-------
Recognizing the problem of Diesel exhaust heavy
hydrocarbon condensation on CVS ducts, sample lines and
sample bags, it had been planned from the outset of the
program to use a heated FID and a heated sample line for
continuous HC analysis, with integration of the FID output
signal giving average HC concentration over the appropriate
test interval. Nominal FID and sample line operating
temperature was 375°F. Strip charts from all tests were
hand integrated by the writer using a planimeter. On a
few tests an electronic counter was used as the integrator
and gave good agreement on mass emissions (within 3%)
with the hand integrations. Unfortunately the counter
proved to be unreliable and could not be used on the
remaining tests.
b. Sample Point
The location of the sample point was determined after
considering these factors: HC. concentrations in the raw
exhaust can vary widely depending on engine operating
mode, and "spikes" during transients are difficult to
integrate. Calculation of mass emissions from the raw
concentrations requires knowledge of engine exhaust
flow rates and is therefore much more difficult than if
a CVS is used. Thus, it was decided to sample from the
dilute exhaust/air mixture.
On the first series of tests the sample was taken at the
same point as the CVS bag, for the sake of expediency.
It soon bee ame apparent that this was not a good sample •
point to use. The first reason is that the partial vacuum
at that point, just upstream of the CVS Roots blower, is
large enough to reduce sample flow to the FID, thereby
reducing response. Secondly, there are several feet of
relatively cold ducting and heat exchanger upstream of
this point where the unburned heavy hydrocarbons can
condense and be lost to analysis.
In general, diesel exhaust should be mixed with
dilution air as close as practical to the vehicle
tailpipe to minimize HC condensation. Also, the sample
for the hot FID should be taken immediately after
thorough mixing has occurred to ensure that a homogeneous
exhaust/air mixture is sampled and to minimize further
HC condensation.
Thus, after the first three tests a different CVS
unit was employed, one in which the exhaust/dilution
air mixing chamber was only about four feet downstream
of the vehicle exhaust pipe. The pitot-type sample probe
was placed in the center of the outlet duct, just inches
downstream of the mixing chamber, and the sample trans-
ported to the hot FID through a heated sample line.
-------
This set up had the effect of reducing response time
of the FID to a change in engine operation from twenty
seconds to four seconds. Traversing the duct with the
probe revealed no appreciable stratification of the mixture,
Beginning with the fifth test (number 18-166), the
pitot probe was replaced with a rake-type probe. Four
equally spaced holes were drilled into a quarter-inch
stainless steel tube in which one end was crimped and
brazed closed. This probe extended diametrically across
the duct with the holes facing upstream.
c. Hot FID Instruments
A Beckman Model 402 hot FID was used on the first
twelve tests and on test No. 15 (18-284), in parallel
with the Scott Model 215 hot FID used on.the last six
tests. Each of these units had a heated sample line as
an integral part of ths system. Operating temperature
was 375°F in both sample line and detector.
I
2. Particulate Emissions • *
The vehicle was taken to Dow Chemical, Midland, Michigan,
for determination of particulate emissions by the Dow procedure,
which includes the cold soak and driving schedule from the 1975
Federal Test Procedure. All the vehicle exhaust is collected
and diluted with air. The samples are then taken from the .
dilute exhaust stream through isokinetic probes and collected
on filters. From the mass of particulates collected on the
filters the mass emissions, in grams per mile, are calculated.
Results and Discussion
1. Gaseous emissions as determined by the 1975 FTP are
presented in Table I in grams per mile (gpm).
a. Hydrocarbons
ThcreThere are two columns of HC data: the normal CVS bag
results (column headed "HC Cold Bag"), and the results
of integrating the continuous hot FID analysis (column
headed "HC Hot FID"). All of the latter data is from the
hand integrations of the strip charts. . On the first six
tests the HC data follow the expected pattern: the hot FID
results are about twice the cold bag results. Table II
shows the results of some sample point investigations which
confirm the approximate 2:1, hot:cold, ratio between the
hot FID analysis at the mixing chamber outlet and the bag
samples.
-------
The next series of tests, numbers 7 through 12, were
run for the Procedures Development Branch and the writer
was not involved in the testing. At the end of this six
test series the writer integrated the hot FID strip charts
and it became apparent that the FID had been malfunctioning
during all these tests. The two sets of HC'data from runs
7 through 12 in Table I show clearly that something was
amiss. The hot FID results should be about twice as high
as the cold bag results, but instead are nearly equal to or
actually lower than the cold bag results. Investigation
of the instrument revealed slow response times and high
background HC levels in the instrument. It was theorized
that the causes were contamination and deterioration of the
instrument over a period of one year of extensive use at
the high temperature conditions in extreme environments.
The unit was thoroughly cleaned and new parts, including
0-rings, gaskets, and filters,were installed.
On the last series of tests, runs 13 through 18, a
Scott Model 215 hot FID was used. The deterioration of
the instrument's performance during this series is clearly
seen in the HC mass data in Tab(le I. Only runs 13 and 14
are considered as valid for HCJdata. The cause for this
gradual decrease in mass emissions data is not known
definitely, but a leak in the sample line is suggested.
After the diesel testing was completed, breakdowns of
this instrument became increasingly frequent and serious,
and included sample pump breakdown, clogging of filters,
and melting of the sample line due to faulty heater control.
b. Carbon Monoxide
With two exceptions, all the carbon monoxide data
listed in Table I were calculated using the concentrations
determined by the most sensitive CO analyzer available in
the laboratory. At least one of two low range instruments
was in use during the Diesel test program. The most
sensitive range of each instrument was 0-200 ppm and
0-250 ppm, respectively. The exceptions are runs 11 and 12.
Inexplicably on these two runs the low range instruments
were not used. Thus, the mass emissions reported (-61 and
.96 gpm, respectively) are considered to be very inaccurate
and should be disregarded. If these two runs are disregarded
the mean CO mass is 1.40 gpm, with a standard deviation of
.09 gpm.
-------
c. Nitrogen Oxides
The NOx data in Table I is quite consistent, with one
test the exception. On run 9 the measured NOx mass was
2.09 gpm, about 27% higher than the all-test mean of 1.64 gpm,
Examination of laboratory records has indicated the
probability that this test was erroneously run at an inertia
weight of 4,500 pounds instead, of the required 3,500 pounds.
The higher inertia weight would explain the increased NOx
emissions as well as the increased C02 emissions on run 9.
If, then, run 9 NOx data is excluded from consideration the •
mean mass is 1.61 gpm, with a standard deviation of .06 gpm,
or less than 4% of the mean. It is felt that NOx emissions
from this vehicle have been well defined.
d. Carbon Dioxide
Carbon dioxide mass emissions are used to calculate
fuel consumption over the Federal test driving schedule,
which simulates an urban-suburban commuting trip. C02
emissions can also be used to determine test validity and
CVS performance. To repeat th$ instance of test number 9,
the high C02 emissions are -an additional indication that
the inertia weight was set too,high.
Test numbers 1,2, and 3 were run on a different dyna-
mometer from all the remaining tests. The step change in
C02 emissions, from about 500 gpm to about 400 gpm, is not
understood. There is confidence that the inertia weights
on the first three tests were correct. It is possible
that the C02 emissions are the result of engine running-in,
since the first three tests were run at odometer readings
of 2060 to about 2100 miles and test number 4 was at
2954 miles.
If tests 1,2,3 and 9 are not included, the mean C02
mass is 410.82 gpm. The standard deviation is 7.77 gpm, or
less than 2% of the mean.
2. Particulate emissions, as determined by the Dow
Chemical procedure, are shown in Table III. These values are
about 7 times higher than those from vehicles using leaded
gasoline, and perhaps 20 times higher than from non-leaded
gasoline. The material was pitch black and very fine.
Conclusions
Mass emissions of all three gaseous pollutants from this
car were at or below the levels required for the 1975 model
year. Average NOx emissions were slightly more than four times
the level required for the 1976 model year. The average C02
emissions were about 50% of C02 emissions from gasoline-powered
cars tested at the same inertia weight.
-------
The 1975 FTP with CVS sampling is applicable for all Diesel
measurements except HC. No difficulties were encountered in the
CVS system as a result of soot build-up.
Hot FID hydrocarbon data is valid from only eight tests: .
1 through 6, 13 and 14. It would be desirable to do further
testing on a Diesel car to define the HC emissions better and to
investigate the effects of lowering FID and sample line
temperatures.
-------
TABLE I
Mercedes-Benz 220 Diesel
Gaseous Emissions
(grams per mile)
1975 Federal Test Procedure
Test No. EPA Test No,
1
2
3
4
5
6
7
8
9*
10
11
12
13
14
15
16
17
18
18-82
18-86
18-87
18-160
18-166
18-172
18-234
18-237
18-242
18-245
18-251
18-255
18-281
18-283
18-284
18-287
18-291
18-292
HC
Cold Bag
.22
.20
.24
.17
.14
.16
.19
.16
.19
.20
,18
.15 •
.15
.15
.14
.14
.14
.14
HC
Hot FID
.31
.37
.42
.38
.31
.30 .
.23
.18
.17
.14
U15
1.16
%34
.28
.21B
.223
.17
.12
.16
CO NOx
1.11 1.61
1.47 1.49
1.56 1.50
1.44 1.47
1.34 1.62
1.37 1.62
1.39 1.62
1.39 1.61
1.38 2.09
1.44 1.67
.61**!. 66
..96**1..67
1.51 1.67
1.40 1.64
1.36 1.61
1.37 1.62
1.39 1.66
1.44 1.62
co2
508.73
495.30
498.21
404.26
398.78
420.16
403.62
413.48
463.88
427.48
413.32
407.47
421.06
416.26
404.73
406.94
407.48
406.42
* 4500 Ibs inertia
**High Range (0-2500 ppm)
B = Beckman 402
S = Scott 215
-------
TABLE II
Sample Point Investigations
1. Vehicle at a steady state cruise condition, approximately
30 mph.
a. Hot FID: 5 ppm (Sample Pt. 2)
Bag Sample: 2.4 ppm (Sample Pt. 4)
b. C02 at Sample Pt. 2: 19.5 meter deflection
C02 at Sample Pt. 3: 19.5 meter deflection
Conclusion: There is no leakage of air into the
flexible duct between mixing chamber and CVS
heat exchanger.
c. Continuous Hot FID analysis
HC at Sample Pt. 2: 5 ppm
HC at Sample Pt. 3: 4.5 ppm
2. Vehicle at a steady state cruise condition.
a. Continuous Hot FID analysis
Sample Pt. 1:. 17 ppm
Sample Pt. 2: 2,2 ppm
Sample Pt. 3: 1.4 ppm
Sample Pt. 4: 1.2 ppm
Sample Point 1 = Vehicle tail pipe (raw exhaust)
2 = Mixing Chamber Outlet Duct (dilute exhaust)
3 = One foot upstream of CVS Heat Exchanger
4 = Bag Sample Point, just upstream of Rootes pump
-------
TABLE III
Mercedes-Benz 220 Diesel
Particulate Emissions
Cgrams per mile)
1975 FTP Driving Schedule
Dow Measuring Procedure
Sample Flow Rate Filter Type Mass, gpm
4 cfm Fiber.glass .6642
1 cfm Fiberglass .7943
1 cfm Fiberglass .7487
1 cfm Anderson + Millipore .7313
------- |